|
|
|
|
|
import pyaudio
|
|
|
from pydub import AudioSegment
|
|
|
import threading
|
|
|
import wave
|
|
|
import time
|
|
|
from pyOpenRPA.Utils import Text
|
|
|
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def DeviceMicrophoneIndex():
|
|
|
"""L-,W+: Выполнить поиск устройства, с помощью которого можно будет выполнить захват c микрофона.
|
|
|
"""
|
|
|
p = pyaudio.PyAudio()
|
|
|
lDeviceInfoDict = p.get_default_input_device_info()
|
|
|
lDefaultIndexInt = lDeviceInfoDict["index"]
|
|
|
return lDefaultIndexInt
|
|
|
|
|
|
def DeviceSystemSoundIndex():
|
|
|
"""L-,W+: Выполнить поиск устройства, с помощью которого можно будет выполнить захват аудио, которое поступает из приложений. Например: аудиоконференции Zoom, whatsapp, telegram и т.д.
|
|
|
"""
|
|
|
p = pyaudio.PyAudio()
|
|
|
inInputBool = True
|
|
|
inIsLoopbackBool = True
|
|
|
if inInputBool == True:
|
|
|
lDeviceInfoDict = p.get_default_output_device_info()
|
|
|
lDefaultIndexInt = lDeviceInfoDict["index"]
|
|
|
lDefaultNameStr = lDeviceInfoDict["name"]
|
|
|
lCatchIndexInt = None
|
|
|
lCatchDiffRatioFloat = 0.0
|
|
|
for lItemDict in DeviceListGet():
|
|
|
lCompareBool = False
|
|
|
if lItemDict["MaxOutputChannelsInt"]>0:
|
|
|
if inIsLoopbackBool==True and lItemDict["HostApiStr"]=="Windows WASAPI": lCompareBool = True
|
|
|
elif inIsLoopbackBool==False: lCompareBool = True
|
|
|
if lCompareBool == True:
|
|
|
lDiffRationFloat = Text.SimilarityNoCase(in1Str=lDefaultNameStr, in2Str=lItemDict["NameStr"])
|
|
|
if lDiffRationFloat> lCatchDiffRatioFloat:
|
|
|
lCatchDiffRatioFloat = lDiffRationFloat
|
|
|
lCatchIndexInt=lItemDict["IndexInt"]
|
|
|
else:
|
|
|
lDeviceInfoDict = p.get_default_output_device_info()
|
|
|
lDefaultIndexInt = lDeviceInfoDict["index"]
|
|
|
lDefaultNameStr = lDeviceInfoDict["name"]
|
|
|
lCatchIndexInt = None
|
|
|
lCatchDiffRatioFloat = 0.0
|
|
|
for lItemDict in DeviceListGet():
|
|
|
lCompareBool = False
|
|
|
if lItemDict["MaxInputChannelsInt"]>0:
|
|
|
if inIsLoopbackBool==True and lItemDict["HostApiStr"]=="Windows WASAPI": lCompareBool = True
|
|
|
elif inIsLoopbackBool==False: lCompareBool = True
|
|
|
if lCompareBool == True:
|
|
|
lDiffRationFloat = Text.SimilarityNoCase(in1Str=lDefaultNameStr, in2Str=lItemDict["NameStr"])
|
|
|
if lDiffRationFloat> lCatchDiffRatioFloat: lCatchIndexInt=lItemDict["IndexInt"]
|
|
|
return lCatchIndexInt
|
|
|
|
|
|
def DeviceListGet():
|
|
|
"""L-,W+: Вернуть список аудио устройст (входящих и исходящих, микрофонов и динамиков).
|
|
|
|
|
|
from pyOpenRPA.Robot import Audio
|
|
|
Audio.DeviceListGet()
|
|
|
|
|
|
:return: [{"IndexInt":1, "NameStr": "",
|
|
|
"HostApiInt": 0, "HostApiStr": "MME"|"Windows WASAPI"|"Windows WDM-KS",
|
|
|
"MaxInputChannelsInt": 0, "MaxOutputChannelsInt": 0,
|
|
|
"DefaultSampleRateFloat": 44100.0
|
|
|
},...]
|
|
|
:rtype: list
|
|
|
"""
|
|
|
l_result = []
|
|
|
p = pyaudio.PyAudio()
|
|
|
for i in range(0, p.get_device_count()):
|
|
|
l_info = p.get_device_info_by_index(i)
|
|
|
l_info_dict = {
|
|
|
"IndexInt":l_info["index"],
|
|
|
"NameStr": l_info["name"],
|
|
|
"MaxInputChannelsInt": l_info["maxInputChannels"],
|
|
|
"MaxOutputChannelsInt": l_info["maxOutputChannels"],
|
|
|
"HostApiInt": l_info["hostApi"],
|
|
|
"DefaultSampleRateFloat": l_info["defaultSampleRate"],
|
|
|
"HostApiStr": p.get_host_api_info_by_index(l_info["hostApi"])["name"] #"MME"|"Windows WASAPI"|"Windows WDM-KS"
|
|
|
}
|
|
|
l_result.append(l_info_dict)
|
|
|
return l_result
|
|
|
|
|
|
class Recorder:
|
|
|
mT=[]
|
|
|
mStatusStr = "0_READY" # "0_READY", "1_RECORDING"
|
|
|
mAudio = None
|
|
|
mCaptureThread = None
|
|
|
mStream = None
|
|
|
|
|
|
mDeviceInt = None
|
|
|
mChannelCountInt = None
|
|
|
mFramesInt = 512
|
|
|
mRecordedFramesList = []
|
|
|
mSampleRateInt = None
|
|
|
mSampleSizeInt = None
|
|
|
|
|
|
mCaptureBool = True
|
|
|
mFolderPathStr = None
|
|
|
mFileNameStr = None
|
|
|
mFileFormatStr = None
|
|
|
mFileAvailableChunkInt = None
|
|
|
mFileNameList=None
|
|
|
|
|
|
mChunkSecFloat = None
|
|
|
mChunkSilentSecFloat = None
|
|
|
|
|
|
mStartSecFloat = None
|
|
|
mStartChunkSecFloat = None
|
|
|
mDurationSecFloat = None
|
|
|
mThresholdInt = 500
|
|
|
|
|
|
mSilentLastCheckTimeFloat = None
|
|
|
|
|
|
mIsMicrophoneBool=None
|
|
|
|
|
|
def __init__(self, inDeviceInt=None):
|
|
|
self.mDeviceInt = inDeviceInt
|
|
|
if inDeviceInt == None: inDeviceInt = DeviceSystemSoundIndex()
|
|
|
self.mDeviceInt = inDeviceInt
|
|
|
if DeviceListGet()[inDeviceInt]["MaxInputChannelsInt"]>0: self.mIsMicrophoneBool = True
|
|
|
else: self.mIsMicrophoneBool = False
|
|
|
|
|
|
def StatusGet(self):
|
|
|
return self.mStatusStr
|
|
|
|
|
|
def CaptureStart(self, inFolderPathStr="",inFileNameStr = "out", inFileFormatStr = "mp3", inDurationSecFloat = None, inChunkSecFloat = 300.0, inChunkSilentSecFloat = 10.0):
|
|
|
# CHECK AUX.mp3
|
|
|
self.mFileNameList=[]
|
|
|
self.mRecordedFramesList=[]
|
|
|
self.mStatusStr = "1_RECORDING"
|
|
|
if inChunkSecFloat != None or inChunkSilentSecFloat != None: self.mFileAvailableChunkInt = 0
|
|
|
self.mDurationSecFloat = inDurationSecFloat
|
|
|
self.mChunkSecFloat = inChunkSecFloat
|
|
|
self.mChunkSilentSecFloat = inChunkSilentSecFloat
|
|
|
self.mSilentLastCheckTimeFloat=time.time()
|
|
|
self.mFolderPathStr = inFolderPathStr
|
|
|
self.mFileNameStr = inFileNameStr
|
|
|
self.mFileFormatStr = inFileFormatStr
|
|
|
self.mAudio = pyaudio.PyAudio()
|
|
|
self.mSampleSizeInt = self.mAudio.get_sample_size(pyaudio.paInt16)
|
|
|
lDeviceInfoDict = self.mAudio.get_device_info_by_index(self.mDeviceInt)
|
|
|
#Open stream
|
|
|
self.mSampleRateInt = int(lDeviceInfoDict["defaultSampleRate"])
|
|
|
self.mChannelCountInt = lDeviceInfoDict["maxInputChannels"] if (lDeviceInfoDict["maxOutputChannels"] < lDeviceInfoDict["maxInputChannels"]) else lDeviceInfoDict["maxOutputChannels"]
|
|
|
self.mStream = self.mAudio.open(format = pyaudio.paInt16,
|
|
|
channels = self.mChannelCountInt,
|
|
|
rate = self.mSampleRateInt,
|
|
|
input = True,
|
|
|
frames_per_buffer = self.mFramesInt,
|
|
|
input_device_index = lDeviceInfoDict["index"],
|
|
|
as_loopback = not self.mIsMicrophoneBool)
|
|
|
self.mCaptureThread = threading.Thread(target=self.__Capture__)
|
|
|
self.mStartSecFloat = time.time()
|
|
|
self.mStartChunkSecFloat = self.mStartSecFloat
|
|
|
self.mCaptureThread.start()
|
|
|
|
|
|
def __Capture__(self):
|
|
|
while self.mCaptureBool==True:
|
|
|
self.mRecordedFramesList.append(self.mStream.read(self.mFramesInt))
|
|
|
self.__TriggerCheck__()
|
|
|
|
|
|
def CaptureStop(self, inWaitStream=True):
|
|
|
self.mCaptureBool=False
|
|
|
if inWaitStream == True: self.mCaptureThread.join()
|
|
|
self.mStream.stop_stream()
|
|
|
self.mStream.close()
|
|
|
#Close module
|
|
|
self.mAudio.terminate()
|
|
|
self.CaptureChunk()
|
|
|
self.mStatusStr = "0_READY"
|
|
|
|
|
|
def CaptureChunk(self):
|
|
|
lFileNameStr = self.mFileNameStr
|
|
|
if self.mFileAvailableChunkInt!=None:
|
|
|
lFileNameStr+=f"_{self.mFileAvailableChunkInt:05}"
|
|
|
self.mFileAvailableChunkInt = self.mFileAvailableChunkInt + 1
|
|
|
# Advanced usage, if you have raw audio data:
|
|
|
sound = AudioSegment(
|
|
|
# raw audio data (bytes)
|
|
|
data=b''.join(self.mRecordedFramesList),
|
|
|
# 2 byte (16 bit) samples
|
|
|
sample_width=self.mSampleSizeInt,
|
|
|
# 44.1 kHz frame rate
|
|
|
frame_rate=self.mSampleRateInt,
|
|
|
# stereo
|
|
|
channels=self.mChannelCountInt
|
|
|
)
|
|
|
sound.export(os.path.join(self.mFolderPathStr,f"{lFileNameStr}.{self.mFileFormatStr}"), format=f"{self.mFileFormatStr}")
|
|
|
self.mFileNameList.append(f"{lFileNameStr}.{self.mFileFormatStr}")
|
|
|
self.mRecordedFramesList = []
|
|
|
self.mStartChunkSecFloat = time.time()
|
|
|
|
|
|
def FileListGet(self):
|
|
|
return self.mFileNameList
|
|
|
|
|
|
def FileLastGet(self):
|
|
|
return self.mFileNameList[-1]
|
|
|
|
|
|
def __Callback__(self, inDefList):
|
|
|
pass
|
|
|
def __CallbackIsSilent__(self):
|
|
|
pass
|
|
|
def __CallbackIsChunked__(self):
|
|
|
pass
|
|
|
def __CallbackIsStopped__(self):
|
|
|
pass
|
|
|
|
|
|
def IsSilent(self, inLastSecFloat=None):
|
|
|
"Returns 'True' if below the 'silent' threshold"
|
|
|
self.mSilentLastCheckTimeFloat = time.time()
|
|
|
if inLastSecFloat == None: inLastSecFloat = self.mChunkSilentSecFloat
|
|
|
lFrameLenInt = int(self.mSampleSizeInt*inLastSecFloat)
|
|
|
if lFrameLenInt<len(self.mRecordedFramesList): lData = self.mRecordedFramesList[-lFrameLenInt:]
|
|
|
else: lData = self.mRecordedFramesList
|
|
|
return max(lData) < self.mThresholdInt
|
|
|
|
|
|
def __TriggerCheck__(self):
|
|
|
"""L-,W+: Контроль записи / остановки аудио по следующим критериям:
|
|
|
- Общая длительность,
|
|
|
- Максимальная длительность части,
|
|
|
- Максимальная длит тишины (часть),
|
|
|
- Максимальная длительность тишины (остановка),
|
|
|
"""
|
|
|
# Проверка по длине записи (CHUNK)
|
|
|
if self.mChunkSecFloat != None and time.time() - self.mStartChunkSecFloat > self.mChunkSecFloat: self.CaptureChunk()
|
|
|
# Остановка записи по максимальной длине
|
|
|
if self.mDurationSecFloat != None and time.time() - self.mStartSecFloat > self.mDurationSecFloat: self.CaptureStop(inWaitStream=False)
|
|
|
# Проверка тишины
|
|
|
#if self.mChunkSilentSecFloat != None and time.time() - self.mSilentLastCheckTimeFloat and self.IsSilent(): self.mT.append("ТИШИНА!!")
|
|
|
|
|
|
|
